Late Health Outcomes Among Survivors of Wilms Tumor Diagnosed Over Three Decades: A Report From the Childhood Cancer Survivor Study

PURPOSE

To evaluate long-term morbidity and mortality among unilateral, nonsyndromic Wilms tumor (WT) survivors according to conventional treatment regimens.

METHODS

Cumulative incidence of late mortality (≥ 5 years from diagnosis) and chronic health conditions (CHCs) were evaluated in WT survivors from the Childhood Cancer Survivor Study. Outcomes were evaluated by treatment, including nephrectomy combined with vincristine and actinomycin D (VA), VA + doxorubicin + abdominal radiotherapy (VAD + ART), VAD + ART + whole lung radiotherapy, or receipt of ≥ 4 chemotherapy agents.

RESULTS

Among 2,008 unilateral WT survivors, 142 deaths occurred (standardized mortality ratio, 2.9, 95% CI, 2.5 to 3.5; 35-year cumulative incidence of death, 7.8%, 95% CI, 6.3 to 9.2). The 35-year cumulative incidence of any grade 3-5 CHC was 34.1% (95% CI, 30.7 to 37.5; rate ratio [RR] compared with siblings 3.0, 95% CI, 2.6 to 3.5). Survivors treated with VA alone had comparable risk for all-cause late mortality relative to the general population (standardized mortality ratio, 1.0; 95% CI, 0.5 to 1.7) and modestly increased risk for grade 3-5 CHCs compared with siblings (RR, 1.5; 95% CI, 1.1 to 2.0), but remained at increased risk for intestinal obstruction (RR, 9.4; 95% CI, 3.9 to 22.2) and kidney failure (RR, 11.9; 95% CI, 4.2 to 33.6). Magnitudes of risk for grade 3-5 CHCs, including intestinal obstruction, kidney failure, premature ovarian insufficiency, and heart failure, increased by treatment group intensity.

CONCLUSION

With approximately 40% of patients with newly diagnosed WT currently treated with VA alone, the burden of late mortality/morbidity in future decades is projected to be lower than that for survivors from earlier eras. Nevertheless, the risk of late effects such as intestinal obstruction and kidney failure was elevated across all treatment groups, and there was a dose-dependent increase in risk for all grade 3-5 CHCs by treatment group intensity.

INTRODUCTION

With advances in risk-adapted, multimodal therapy, 85%-90% of children diagnosed with Wilms tumor (WT) are expected to become 5-year survivors.^1,2 WT has traditionally been treated with nephrectomy + vincristine and actinomycin D (VA)–based chemotherapy, with addition of doxorubicin and radiotherapy (RT) for advanced locoregional or metastatic disease.^3,4 Risk stratification of WT therapy has sought to achieve high survival rates, while limiting exposure to therapies that increase risk for late effects, which has resulted in an overall reduction in therapy for lower-risk patients. Doxorubicin exposure was reduced by limiting its use to more advanced disease, along with reductions in the cumulative dose. Additional modifications in WT therapy include limiting use of flank and/or whole abdominal radiotherapy (ART) to those with advanced locoregional disease, reducing total dose of RT delivered, omitting whole lung radiotherapy (WLRT) for lung metastases in patients exhibiting complete response to chemotherapy, and identification of patients where nephrectomy alone may provide sufficient treatment.^5-9

CONTEXT

• Key Objective

• To determine the risks for long-term morbidity and mortality experienced by survivors of unilateral, nonsyndromic Wilms tumor (WT) and characterize how risks for late effects are affected by commonly used WT treatment regimens.

• Knowledge Generated

• WT survivors receiving treatment limited to nephrectomy, vincristine, and actinomycin-D experience rates of late chronic health conditions and mortality that are largely comparable with noncancer populations. Survivors treated with more intense regimens experience elevated rates of late mortality and late health conditions such as heart failure, premature ovarian insufficiency, and subsequent malignant neoplasms. Risk for these conditions is further influenced, in a dose-dependent manner, by exposure to doxorubicin and/or radiotherapy, and is most pronounced among survivors treated for high-risk disease, including those with early (< 5 years) WT relapse.

• Relevance (S. Bhatia)

• This study informs the need for risk-based anticipatory follow-up and screening for WT patients at high risk for long-term complications.*

  *Relevance section written by JCO Associate Editor Smita Bhatia, MPH, MD, FASCO.

There is a paucity of information describing temporal changes in the burden of long-term morbidity and mortality experienced by WT survivors as they age. Although it has been speculated that late toxicities are likely modest,^10,11 relatively little is known about the late effects associated with common curative approaches such as use of nephrectomy with VA for localized disease, as compared with more intensive approaches using doxorubicin and RT for more advanced disease, where toxicities have been well delineated. Furthermore, research is needed among WT survivors with extended follow-up to capture important late-occurring health outcomes, such as kidney failure, that may develop as survivors age into later adulthood.^12-14 We leveraged the Childhood Cancer Survivor Study (CCSS) cohort to evaluate late mortality, chronic health conditions (CHCs), and self-reported health-related quality of life (HRQOL) in a large cohort of WT survivors diagnosed over three decades to help address these gaps in knowledge.

METHODS

Population

The CCSS is a multi-institutional, retrospectively constructed cohort study with a longitudinal prospective follow-up of 25,665 5-year survivors of childhood cancer diagnosed between 1970 and 1999.^15,16 The current analysis is limited to survivors of unilateral, nonsyndromic WT (n = 2,008) and sibling controls (n = 5,006 siblings of CCSS survivors, not limited to WT survivors). Survivors were grouped by decade so that trends in therapy could be visually inspected and described, but these trends were not formally tested statistically. Details regarding selection of the population are provided in Appendix 1 (online only) and Appendix Figure A1 (online only).

Outcomes and Covariates

Late mortality was defined as death occurring after 5 years from original WT diagnosis. Deaths were identified by linkage with the National Death Index through December 2017. The International Classification of Disease—9th and 10th Revision was used to categorize the underlying cause of death using previously established rules.¹⁷ Late deaths were further categorized as because of WT relapse, other health-related causes (all health-related deaths, including those from late effects), or external (eg, accident and injury) causes. For analyses examining mortality because of neoplasm, secondary neoplasms (includes nonmalignant neoplasms for this analysis only) were assessed to facilitate general population comparisons given that mortality data limited specifically to malignant neoplasm are not available for the general population.

Subsequent malignant neoplasms (SMNs) were defined as new malignant neoplasms (excluding nonmelanoma skin cancer and meningioma), occurring 5 or more years after WT diagnosis. SMNs were ascertained by self-report and verified by medical record review.

Self-reported CHCs were graded according to Common Terminology Criteria for Adverse Events (CTCAE, version 4.03) and categorized as mild (grade 1), moderate (grade 2), severe or disabling (grade 3), life-threatening (grade 4), or fatal (grade 5).¹⁸ Outcome-specific analyses were limited to the most commonly reported grade 3-5 CHCs, including intestinal obstruction requiring operative intervention,¹⁹ kidney failure requiring dialysis or kidney transplant,²⁰ heart failure,²¹ and gonadal dysfunction.²² Because of very few outcomes of gonadal dysfunction among males (3 of 893 male survivors), analysis of gonadal dysfunction was limited to premature ovarian insufficiency (POI) in females, defined as cessation of ovarian function before age 40 years. The Short Form-36 was used to measure HRQOL.²³ Survey responses were reported and dichotomized into impaired (v not impaired) using 16th percentile corresponding to scores that are one standard deviation below or lower compared with population normative data.²⁴

Covariates included demographic data and treatment exposures abstracted under uniform protocols from medical records at treating institutions after family or participant consent. Chemotherapeutic agents were evaluated on the basis of established WT regimens.²⁵ Survivors were classified as having received RT (yes/no), and we estimated delivered doses to the most common treatment sites using RT records: (1) ipsilateral flank RT, (2) whole abdomen radiotherapy (WART), and (3) WLRT.²⁶ Exposure to RT in the entire WT cohort was evaluated in two types of analyses: by RT site and by RT dose. For RT site analysis, survivors were stratified (yes/no) for two field categories, WLRT and/or ART, which included both the ipsilateral flank and whole abdomen fields. For RT treatment dose analysis, flank field RT, WART and WLRT were categorized as > 20 Gy and ≤ 20 Gy.

A subset of WT survivors (n = 1,261; 62.8% of the cohort) were classified according to commonly used treatment regimens (in order of lowest to highest intensity): (1) nephrectomy and VA (n = 623); (2) nephrectomy, VA, doxorubicin, and ART only (VAD + ART only, n = 396); (3) nephrectomy, VAD, ART, and WLRT (VAD + ART + WLRT; n = 91); and (4) nephrectomy and four or more chemotherapy agents with any RT (n = 151), typically reserved for patients with more unfavorable histologic features, poor initial response, and/or relapse.²⁷ Additional analyses comparing outcomes between survivors who could and could not be categorized into treatment groups were conducted to elucidate any possible bias introduced by the chosen classification scheme. WT survivors were also stratified by those who did and did not experience early relapse, defined as recurrence of WT within 5 years of initial diagnosis, to facilitate a limited outcome comparison between these groups.

Statistical Analysis

Follow-up at-risk time began at 5 years after original WT diagnosis for survivors; for siblings, cohort entry occurred on the date of their corresponding survivors. Follow-up at-risk time for mortality analysis ended on participant death or December 31, 2017, the last date of the National Death Index search. Follow-up at-risk time for nonmortality analysis ended on occurrence of the event of interest, participant death (competing risk event), or the completion of the most recent questionnaire.

Cumulative incidence of each outcome was calculated. Standardized mortality ratios (SMRs) were used to quantify the post–5-year mortality rate among WT survivors compared with rates from an age-, sex-, and calendar-year-matched US population. For SMNs, standardized incidence ratios (SIRs) for survivors were calculated using age-, sex- and calendar-year-specific cancer incidence rates from the Surveillance, Epidemiology and End Results program.² Adjusted rate ratios (RRs) and 95% CIs evaluating associations between individual treatment exposures or treatment groups with outcomes were estimated using multivariable piecewise-exponential models adjusting for sex, race, and attained age as cubic splines (knots: 15, 25, 30, 35, and 45). For analyses examining association between treatment exposures and specific CHCs, covariates included chemotherapy and RT fields and doses as earlier described. Multivariable logistic regression was used to compare associations of self-assessed HRQOL across treatment groups adjusting for age, sex, and race with odds ratios (ORs), 95% CI, and P values reported. Generalized estimating equations were used to account for potential within-family correlation for survivor-sibling comparisons.

All tests were two-sided with P < .05 considered statistically significant. SAS (version 9.4, SAS Institute Inc., Cary, NC) and R Statistical Software (version 4.0.5) were used for analyses.

RESULTS

Demographic and Treatment Characteristics

Characteristics of the 2,008 unilateral WT survivors and 5,006 siblings are summarized in Table 1. The median age of survivors at diagnosis and follow-up were 3.2 years (range, 0-20.2 years) and 27.8 years (range, 6.4-57.5 years), respectively. All survivors underwent unilateral nephrectomy, and nearly all (98.1%) received chemotherapy, with the most common exposures being vincristine (93.8%), actinomycin D (93.5%), and doxorubicin (44.8%). Fifty-five percent of survivors received RT, with the most common fields being flank (27.8%), whole abdomen (13.8%), and whole lung (13.4%). The proportion of survivors receiving any RT declined from the 1970s to the 1990s. This was accompanied by a reduction of RT dose throughout the decades. Survivors were more likely to be treated with doxorubicin over time, but the proportions of survivors treated with cumulative dose of ≥ 250 mg/m² decreased over time. Use of other agents, including etoposide, cyclophosphamide, and platinum agents, was also more common in the 1990s compared with the 1970s and 1980s. Among 2,008 survivors, 68 (3.4%) experienced early WT relapse. The ≥ 4 drugs + any RT treatment group included the greatest proportion of survivors experiencing early relapse (23.2%) relative to the other treatment groups (0%-3%; data not shown).

TABLE 1.

Demographic and Treatment Characteristics of Wilms Tumor Survivors and Siblings

Late Mortality

There were 142 late deaths, with the 35-year cumulative incidence of all-cause, relapse-related, and health-related late mortality of 7.8%, 1.6%, and 4.7%, respectively (Fig 1). The most frequent causes of death were SMN (n = 42), WT relapse (n = 30), and cardiac (n = 9; Appendix Table A1, online only). Compared with the general population, survivors had an increased rate of all-cause (SMR, 2.9; 95% CI, 2.5 to 3.5) and health-related (SMR, 5.4; 95% CI, 4.4 to 6.4) mortalities (Table 2). For survivors treated with VA, both all-cause (SMR, 1.0; 95% CI, 0.5 to 1.7) and health-related (SMR, 1.5; 95% CI, 0.6 to 3.0) late mortality rates were statistically comparable with those of the general population. Risk for overall and health-related mortality was elevated for all other treatment groups, especially those treated with ≥ 4 chemotherapy agents + RT.

FIG 1.

Cumulative incidence (%) of late mortality among WT survivors overall, by treatment groups, and according to relapse status: (A) all-cause mortality, (B) all-cause mortality,^a (C) cause-specific mortality,^a (D) health-related mortality, (E) health-related mortality,^a and (F) SN-cause mortality. ^aIncludes survivors not classified into one of the four treatment groups. ART, abdominal radiotherapy; RT, radiotherapy; SN, secondary neoplasm (may include death because of nonmalignant neoplasms); VA, vincristine and actinomycin D; VAD, vincristine, actinomycin D, doxorubicin; WLRT, whole lung radiotherapy; WT, Wilms tumor.

TABLE 2.

Late Mortality, SMNs, and CHCs Among all Unilateral WT Survivors and by Treatment Regimen

Late Morbidity

Overall WT survivor cohort.

The 35-year cumulative incidence of any grade 3-5 CHC was 34.1% (95% CI, 30.7 to 37.5) among WT survivors and 14.8% (95% CI, 13.4 to 16.2) among siblings (RR, 3.0; 95% CI, 2.6 to 3.5; Fig 2, Table 2, and Appendix Table A1). Seventy-three survivors developed 82 SMNs (SIR, 4.1; 95% CI, 3.2 to 5.1), representing a 35-year cumulative incidence of 6.1% (95% CI, 4.3 to 8.0). Breast (n = 21; SIR, 6.9; 95% CI, 3.9 to 11.2), thyroid (n = 10; SIR, 4.7; 95% CI, 2.3 to 8.4), and intestinal/colorectal (n = 10; SIR, 12.0; 95% CI, 5.5 to 22.7) were the most frequently reported SMNs. Among all survivors, the 35-year cumulative incidences of selected CHCs were 8.1% (95% CI, 6.6 to 9.6) for intestinal obstruction, 7.3% (95% CI, 5.0 to 9.7) for POI, 4.0% (95% CI, 2.6 to 5.5) for heart failure, and 2.4% (95% CI, 1.6 to 3.2) for kidney failure, providing estimated risks relative to siblings with RRs of 16.2 (95% CI, 9.2 to 28.4), 4.2 (95% CI, 2.7 to 6.5), 11.8 (95% CI, 6.3 to 22.1), and 10.4 (95% CI, 5.0 to 21.8), respectively.

FIG 2.

Cumulative incidence (%) of grade 3-5 chronic health conditions among WT survivors according to the treatment group and relapse status: (A) grade 3−5 CHCs, (B) grade 3−5 CHCs,^a (C) ≥ 2 grade 3−5 CHCs, (D) ≥ 2 grade 3−5 CHCs,^a (E) heart failure, (F) heart failure,^a (G) kidney failure, (H) kidney failure,^a (I) intestinal obstruction, (J) intestinal obstruction,^a (K) premature ovarian insufficiency, and (L) premature ovarian insufficiency.^{a a}Includes survivors not classified into one of the four treatment groups. ART, abdominal radiotherapy; CHC, chronic health conditions; RT, radiotherapy; VA, vincristine and actinomycin D; VAD, vincristine, actinomycin D, doxorubicin; WLRT, whole lung radiotherapy; WT, Wilms tumor.

WT treatment groups.

Compared with siblings, WT survivors treated with VA experienced a modestly increased rate of developing any grade 3-5 CHC (RR, 1.5; 95% CI, 1.1 to 2.0), but there was no increase in the rate of developing heart failure or POI. Relative to siblings, survivors treated with VA remained at elevated risk for intestinal obstruction (RR, 9.4; 95% CI, 3.9 to 22.2) and kidney failure (RR, 11.9; 95% CI, 4.2 to 33.6), but the magnitude of risk was lower than that for more intensive treatment groups (Table 2). Survivors experiencing early relapse exhibited an increased cumulative incidence of kidney failure compared with those who did not (9.1%, 95% CI, 3.6 to 17.5 v 2.2%, 95% CI, 1.5 to 3.1), with more prevalent events at 5 years (Fig 2 and Appendix Table A1). WT survivors treated with more intense therapy generally experienced greater rates of grade 3-5 CHCs, heart failure, and POI compared with WT survivors treated with VA treatment alone (Appendix Table A2, online only).

Additional analyses comparing mortality and morbidity between survivors who could and could not be categorized into treatment groups are summarized in Appendix Table A3 (online only). Although estimates for SMR, SIR for SMNs, and RR for CHCs varied slightly across groups, the strengths and patterns of association appear similar with the exception of a higher rate of intestinal/colorectal SMNs among the survivors who could not be classified.

Specific treatment exposures.

WLRT was associated with increased risk of SMN, heart failure, and intestinal obstruction compared with survivors receiving no RT (Table 3). WART was associated with increased risk of SMN, heart failure, intestinal obstruction, and POI, and the magnitudes of risk increased with dose. Flank RT > 20 Gy was associated only with an increased risk of intestinal obstruction compared with no RT, whereas flank RT ≤ 20 Gy was not associated with SMN or any other CHC.

TABLE 3.

Multivariable-Adjusted RRs for Chronic Health Conditions in Unilateral WT Survivors by Specific Treatment Exposure

Self-Reported HRQOL

Overall, survivors were more likely to report poor general health (OR, 2.4; 95% CI, 1.8 to 3.1) and poor physical functioning (OR, 2.9; 95% CI, 1.8 to 4.6) HRQOL compared with siblings (Table 4). Survivors were also more likely to have increased impairments in mental domains of HRQOL including more fatigue and emotional and social role limitations. Increased risk of poor outcomes was observed slightly more frequently for more intense treatment regimens, particularly within physical domains, but there was not a strong pattern in the relationship between WT treatment groups and HRQOL. Survivors treated with VA reported comparable HRQOL as siblings.

TABLE 4.

Self-Assessed Physical, Mental, and Neurocognitive HRQOL Among All WT Survivors and According to the Treatment Group Compared With Siblings

DISCUSSION

We report results from analyses of 2,008 long-term unilateral WT survivors, representing to our knowledge the largest study to include outcomes according to common treatment regimens. For the overall cohort, the burden of late nonrelapse mortality and morbidity was substantial, increased sharply with longer follow-up, and was associated with more intense therapies. On the other hand, survivors who were treated with VA did not exhibit a statistically significant increased risk for overall mortality, experienced minimal-to-no increase in health-related mortality, and had no increased risk for important CHCs such as heart failure and POI, highlighting the expected low burden of late effects in this population. The findings from the current study could inform future treatment for WT and guide needed risk-based screening and life-long follow-up to minimize the evolving burden of late effects for this growing population of long-term survivors and on health care systems.¹⁸

Nearly all patients with WT undergo nephrectomy. Therefore, the risk for late renal dysfunction is a unique concern.^28-30 In the current study, the 35-year cumulative incidence of late kidney failure among survivors was 2.4%, an approximately 10-fold increased rate relative to siblings. No single chemotherapy or RT-related exposure was associated with the outcome, suggesting that nephrectomy may be the primary risk factor. Although prior reports describing late kidney failure among unilateral WT survivors are sparse, the 20-year rate has been estimated to be 0.6% on the basis of data from the National Wilms Tumor Study.¹³ In our study, the 20-year rate was 1.7%, a higher risk that could be due to inclusion of more survivors who might have been treated off clinical trials and/or before widespread treatment standardization. Although survivors experiencing early relapse exhibited a higher 35-year cumulative incidence of kidney failure compared with those who did not, the overall incidence of kidney failure did not substantially change when relapsed survivors were excluded. Therefore, the increased incidence compared with prior reports likely cannot be attributed solely to the inclusion of relapsed survivors. The steady increase in incidence suggests that kidney failure may be an ongoing risk for WT survivors living beyond reported follow-up times in previous studies. Accordingly, survivors should be counseled about their increasing risk for kidney failure with age and the importance of screening and management of comorbidities (eg, hypertension and diabetes) that would place them at elevated risk. Efforts to investigate the role of nephron-sparing surgery in unilateral patients with WT may also be justified under carefully designed clinical trials.^31,32

The combination of VA and unilateral nephrectomy is currently the standard of care for the management of most stage I and II WT in North America, comprising approximately 40% of newly diagnosed patients. The current study highlights several reassuring findings for these survivors, including no increased risk of overall mortality and no significantly elevated rate of heart failure and POI, likely owing to the omission of doxorubicin and/or RT in this group.³³ The increased rate of intestinal obstruction likely reflects the late morbidity associated with major abdominal surgery such as nephrectomy.^19,34 Moving forward, it will be important to determine whether evolving treatment strategies (eg, omission of chemotherapy for select patients with WT or consideration of minimally invasive surgery) could further reduce late health consequences while preserving recurrence-free survival, particularly given the increased risk for late mortality and other poor health outcomes associated with early disease relapse in this cohort.^5,9

Consistent with reports from other survivor populations, we observed increases in rates of late morbidity and mortality among survivors treated with doxorubicin and/or RT. High-dose (≥ 250 mg/m², cumulative dose) doxorubicin exposure was associated with a nearly five-fold rate of development of heart failure, compared with no doxorubicin exposure. The rate of intestinal obstruction was highest among survivors treated with higher-dose (> 20 Gy) whole abdomen or flank RT. These findings speak to an additive risk of RT that is beyond the risk of intestinal obstruction because of nephrectomy alone.¹⁹ We also found an increased rate of POI associated with the receipt of ART that appeared to be dose-dependent providing justification for discussions focused on fertility preservation options for female patients with WT requiring whole ART.^35,36 Despite its established association with gonadotoxicity, we observed no association between cyclophosphamide exposure and POI and speculate that this is likely due to small numbers of survivors treated with this agent, especially at gonadotoxic doses, in this cohort and possibly a more dominant role of ART and higher doses of doxorubicin in the development of this outcome. In addition, the low rate of self-reported gonadal dysfunction among males may be due to the disparate impact of gonadotoxic therapies on Leydig cell compared with Sertoli cell function, such that male survivors may progress through puberty and experience normal sexual function despite experiencing infertility, which may be undiagnosed and under-reported.³⁷ To limit overall morbidity for WT survivors, a major initiative in the development of modern treatment protocols has been the reduction of anthracycline and/or radiation dose and omission of radiotherapy for certain populations.^14,38-43 The findings from the current study support the need for continuous vigilance in these efforts and reinforce the importance of avoidance of tumor spillage during surgery or inadequate surgical staging that may result in disease upstaging.^6,39,44

Several limitations with the present study should be considered. First, all nonmortality outcomes in the CCSS cohort are self-reported. Considering this, we were careful to focus on health complications that were less likely to be under-reported owing to their severity. Still, the results of our study may under-represent the true burden of CHCs for this population because of under-reporting.^45-48 In addition, given that survival for children with WT has continued to improve with time, it is possible that the incidence of CHCs in more contemporary cohorts will be higher than those reported in the current study. Next, although we attempted to exclude them from analysis, we acknowledge that identification of patients with known WT predisposition syndromes—for which rates of kidney failure and other CHCs may be higher—might have been historically difficult and some might have been inadvertently included. Finally, because not all survivors could be categorized into one of the four treatment groups, this represents a potential source of bias that could affect interpretation of these comparisons. Nevertheless, we feel that this bias is likely minimal on the basis of the data reported in Appendix Table A3. Comparisons between treatment groups are also limited because of the way in which treatment is differentially prescribed by stage of disease and has changed over multiple decades. Treatment-related decisions must continue to prioritize cure while accounting for long-term implications of treatment rendered.

In conclusion, many WT survivors experience elevated rates of late morbidity and mortality, especially as treatment is intensified. Survivors treated with the VA regimen, on the other hand, experience low rates of overall and health-related late mortality and important late CHCs such as heart failure and POI and report HRQOL that is comparable with sibling controls. These findings provide important evidence to suggest that risk-adapted therapy for the treatment of WT can successfully maintain high disease survival rates while optimizing the long-term health of survivors. Rate of late kidney failure increases with age and remains a unique concern among all WT survivors necessitating careful life-long follow-up.

APPENDIX 1. SUPPLEMENTARY METHODS

Two thousand two hundred seventy-four Wilms survivors were identified within the Childhood Cancer Survivor Study (CCSS) cohort: 1,256 from the original cohort (diagnosed 1970-1986) and 1,018 from the expansion cohort (diagnosed 1987-1999).

A cohort consisting of survivors of unilateral Wilms tumor (WT) without known WT predisposition syndromes was generated by excluding, first, survivors with bilateral WT and, second, survivors with a high likelihood of having a known WT predisposition syndrome on the basis of responses to CCSS survey questions (Appendix Fig A1). Bilateral status was known among survivors in the expansion cohort (n = 70). Within the original cohort, bilateral status was inferred on the basis of operative procedures performed at the time of diagnosis. Survivors undergoing the following procedures or combination of procedures were considered highly likely to have been treated for bilateral tumors and/or have a known WT predisposition syndrome and were excluded (n = 85):

• 55.4 (partial nephrectomy)

• 55.54 (bilateral complete nephrectomies)

• 55.5 on two occasions (bilateral complete nephrectomies on separate occasions)

• 55.52 (nephrectomy of remaining kidney)

Responses to CCSS survey questions were used to exclude survivors considered very likely to have a WT predisposition syndrome. Survivors and siblings who responded in an affirmative way to having any of the following conditions were excluded:

• Beckwith-Wiedemann Syndrome, Bloom Syndrome

• Fanconi anemia

• Blindness at birth/different colored eyes or missing an iris (aniridia)

• Unequal limb length at birth; genital anomalies; undescended testicles

As a result of these criteria, an additional 111 WT survivors and 39 siblings were excluded, resulting in a final cohort of 2,008 WT survivors and 5,006 siblings.

For analyses by treatment group, an additional 747 Wilms survivors were excluded because they did not receive treatment according to the four most common treatment regimens. As such, analyses including all WT survivors included all 2,008 survivors as derived above, and analyses by treatment group included only the 1,261 survivors (62.8% of the full cohort) who could be classified accordingly. Reasons that survivors were unable to be classified according to one of the four treatment groups are as follows:

n = 30 treated with surgery only; n = 47 treated with vincristine, actinomycin D, doxorubicin, and no radiotherapy or radiotherapy data missing; n = 257 treated with vincristine and actinomycin D and with radiotherapy or radiotherapy data missing; n = 92 treated with four or more chemotherapy agents but no radiotherapy; n = 124 encompassing a variety of reasons, for example. chemotherapy regimens not including vincristine, actinomycin D, and doxorubicin, one chemotherapy agent only, or surgery and radiotherapy only.

TABLE A1.

35-Year Cumulative Incidences (%) of Mortality and CHCs for All WT Survivors, Siblings, and WT Survivors According to the Treatment Group

TABLE A2.

CHCs Among WT Survivors According to the Treatment Group

TABLE A3.

Late Mortality, SMNs, and CHCs Among All Unilateral WT Survivors and Those Who Could and Could Not Be Categorized Into One of the Four Treatment Groups

FIG A1.

Derivation of the final cohort. CCSS, Childhood Cancer Survivor Study; LTFU, lost to follow-up; WT, Wilms tumor.

AUTHOR CONTRIBUTIONS

Conception and design: Brent R. Weil, Christopher B. Weldon, Elizabeth A. Mullen, Arin L. Madenci, Wendy M. Leisenring, Kevin C. Oeffinger, Amanda M. Termuhlen, Leslie L. Robison, Gregory T. Armstrong, Eric J. Chow, Saro H. Armenian

Financial support: Gregory T. Armstrong

Administrative support: Christopher B. Weldon, Leslie L. Robison, Gregory T. Armstrong

Provision of study materials or patients: Joseph P. Neglia, Amanda M. Termuhlen, Gregory T. Armstrong

Collection and assembly of data: Brent R. Weil, Rebecca M. Howell, Susan A. Smith, Elizabeth A. Mullen, Wendy M. Leisenring, Joseph P. Neglia, Lucie M. Turcotte, Leslie L. Robison, Gregory T. Armstrong

Data analysis and interpretation: Brent R. Weil, Andrew J. Murphy, Qi Liu, Christopher B. Weldon, Elizabeth A. Mullen, Arin L. Madenci, Kevin C. Oeffinger, Amanda M. Termuhlen, Sogol Mostoufi-Moab, Jennifer M. Levine, Kevin R. Krull, Yutaka Yasui, Gregory T. Armstrong, Eric J. Chow, Saro H. Armenian

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Late Health Outcomes Among Survivors of Wilms Tumor Diagnosed Over Three Decades: A Report From the Childhood Cancer Survivor Study

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).